Rates are per 100 000 and age adjusted to the 2000 US standard population for localized disease (A), regional disease (B), and distant disease (C). The trend lines are logarithmic. APC indicates annual percentage change.
Bailey CE, Hu C, You YN, Bednarski BK, Rodriguez-Bigas MA, Skibber JM, Cantor SB, Chang GJ. Increasing Disparities in the Age-Related Incidences of Colon and Rectal Cancers in the United States, 1975-2010. JAMA Surg. 2015;150(1):17-22. doi:10.1001/jamasurg.2014.1756
The overall incidence of colorectal cancer (CRC) has been decreasing since 1998 but there has been an apparent increase in the incidence of CRC in young adults.
To evaluate age-related disparities in secular trends in CRC incidence in the United States.
Design, Setting, and Patients
A retrospective cohort study using the Surveillance, Epidemiology, and End Results (SEER) CRC registry. Age at diagnosis was analyzed in 15-year intervals starting at the age of 20 years. SEER*Stat was used to obtain the annual cancer incidence rates, annual percentage change, and corresponding P values for the secular trends. Data were obtained from the National Cancer Institute’s SEER registry for all patients diagnosed as having colon or rectal cancer from January 1, 1975, through December 31, 2010 (N = 393 241).
Main Outcome and Measure
Difference in CRC incidence by age.
The overall age-adjusted CRC incidence rate decreased by 0.92% (95% CI, −1.14 to −0.70) between 1975 and 2010. There has been a steady decline in the incidence of CRC in patients age 50 years or older, but the opposite trend has been observed for young adults. For patients 20 to 34 years, the incidence rates of localized, regional, and distant colon and rectal cancers have increased. An increasing incidence rate was also observed for patients with rectal cancer aged 35 to 49 years. Based on current trends, in 2030, the incidence rates for colon and rectal cancers will increase by 90.0% and 124.2%, respectively, for patients 20 to 34 years and by 27.7% and 46.0%, respectively, for patients 35 to 49 years.
Conclusions and Relevance
There has been a significant increase in the incidence of CRC diagnosed in young adults, with a decline in older patients. Further studies are needed to determine the cause for these trends and identify potential preventive and early detection strategies.
Quiz Ref IDColorectal cancer (CRC) is the third most common cancer among men and women, with an estimated 142 820 new cases and an estimated 50 830 deaths in the United States in 2013.1 Both the incidence and mortality rates of CRC have been decreasing in the United States.1 From 1998 through 2006, the incidence of CRC has decreased 3.0% per year in men and 2.4% per year in women.2 The observed decline in incidence is largely attributed to an increase in screening, specifically colonoscopy, which is recommended for all adults 50 years or older.2- 5 Screening reduces the incidence of CRC by detecting and removing adenomatous polyps.6,7 Conversely, the incidence of CRC in adults younger than 50 years, for whom CRC screening is not recommended, appears to be increasing, and these patients are more likely to present with advanced disease.8- 10 Although no clear reasons have been identified, similar results have been observed in young women with breast cancer.11 Based on these observations, the aim of this study was to perform a population-based evaluation of age-related disparities in secular trends in CRC incidence in the United States.
We performed a retrospective cohort study of patients with histologically confirmed colon or rectal cancer using the Surveillance, Epidemiology, and End Results (SEER) database. Institutional review board approval was not obtained for the present study as it was not applicable.
Quiz Ref IDThe SEER database is an authoritative source for cancer incidence, survival, and prevalence, currently representing approximately 28% of the US population. The SEER program collects demographic information (eg, age, sex, and race/ethnicity) and clinical information (eg, primary tumor site, tumor histology, tumor grade, stage, treatment, and survival) from 17 cancer registries across the United States. To evaluate secular trends beyond recent years, we used cases identified in SEER 9, which includes San Francisco–Oakland, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle–Puget Sound, Utah, and Atlanta (metropolitan).12 Because the Seattle–Puget Sound and Atlanta registries joined the SEER program in 1974 and 1975, respectively, we selected the 1975 through 2010 cohort for analysis. Cases reported from nursing homes, autopsies, and death certificates were excluded.
The annual cancer incidence rates, annual percentage changes (APCs), and corresponding P values of trends were determined using SEER*Stat (version 8.04; National Cancer Institute). The APC indicates the cancer rate change at a constant percentage of the rate of the previous year and was calculated by fitting a weighted-least-squares regression line to the natural logarithm of the rates. The APC is commonly used to characterize trends in cancer rates over time. We used the widely accepted built-in function of SEER*Stat, which precluded model diagnostics. To minimize the effect of a difference in age distributions over time, all rates were age adjusted to the 2000 US standard population for trend analysis. All P values for significance testing of APC = 0 were 2 sided and considered to be of statistical significance when P < .05.
The predicted annual incidence rates were calculated based on the estimated APC by age and cancer site and assumed to change at a constant percentage of the rate of the previous year. For example, if the APC was 1% and the rate was 10 000 per 100 000 in 2000, the predicted rate would be 10 000 × 1.01 = 10 100 per 100 000 in 2001 and 10 100 × 1.01 = 10 201 per 100 000 in 2002.
We analyzed age at diagnosis in 15-year intervals starting at the age of 20 years. We also used the SEER historic stage A definition for extent of disease at diagnosis. The SEER historic stage A variable is a simplified version of stage derived from Extent of Disease from 1973-2003 and Collaborative Stage for 2004 and beyond. The historic staging variable has the advantage of being the only staging variable that has been recorded consistently over the study period. Localized disease refers to disease confined to the colon or rectum, regional disease refers to contiguous and adjacent organ spread (eg, lymph nodes, kidney, and pelvic wall), and distant disease refers to remote metastases.
There were 393 241 patients with histologically confirmed colon or rectal cancer who met inclusion criteria and were analyzed. A total of 281 462 patients (71.6%) had colon cancer, of which 38.8% presented with right (including transverse) colon cancer, 30.9% presented with left colon cancer, and 1.9% presented with disease not otherwise specified. The remaining patients (n = 111 779) presented with rectosigmoid and rectal cancers. Overall, the age-adjusted CRC incidence rates significantly decreased between 1975 and 2010 (APC, −0.92; 95% CI, −1.14 to −0.70; P < .001). Results by sex, race/ethnicity, age, stage at diagnosis, site of disease, and geographic region are shown in the Table. Quiz Ref IDThe CRC incidence rates declined overall by 1.03% in men and 0.91% in women. The rates for white, black, and other individuals decreased by less than 1%. The most pronounced decline was observed in patients aged 75 years or older (APC, −1.15; 95% CI, −1.47 to −0.84; P < .001). However, the CRC incidence rates increased for patients aged 20 to 49 years, with the most pronounced increase observed in patients aged 20 to 34 years (APC, 1.99; 95% CI, 1.45-2.51; P < .001). Other factors associated with decreasing CRC incidence rates included stage, site of disease, and geographic region.
The annual incidence rates of colon cancer from 1975 to 2010 by 15-year intervals and stage of disease at diagnosis are shown in Figure 1. Since 1975, there has been a steady decline in the incidence rate of CRC in patients aged 50 years and older for each stage category (ie, localized, regional, and distant). The most notable declines were in regional and distant disease. For regional disease, the APCs for patients aged 50 to 74 years and patients older than 75 years were −1.41 (95% CI, −1.75 to −1.07; P < .001) and −1.46 (95% CI, −1.86 to −1.07; P < .001), respectively. For patients with distant disease, the APCs for patients aged 50 to 74 years and patients older than 75 years were −1.20 (95% CI, −1.35 to −1.05; P < .001) and −1.53 (95% CI, −1.79 to −1.28; P < .001), respectively. The incidence rates of localized and regional disease in patients 35 to 49 years also decreased but there was a nonsignificant increase in the incidence rate of distant disease in this age group. Opposite trends were observed in patients younger than 34 years when compared with patients older than 50 years. The annual incidence rate for colon cancer diagnosed in this group of patients is increasing across all 3 stage groups. The APCs for patients with localized, regional, and distant disease at diagnosis in the 20- to 34-year age group were 1.10 (95% CI, 0.28-1.93; P = .01), 0.91 (95% CI, 0.23-1.61; P = .01), and 1.81 (95% CI, 0.88-2.75; P < .001), respectively.
The APC-based predicted incidence rate of colon cancer in 2020 and 2030 compared with 2010 for the 4 different age groups are shown in Figure 2. Based on the predictive model, the incidence rate of colon cancer will continue to decrease for patients older than 50 years. The largest decrease is expected for patients older than 75 years. The opposite trend is noted in patients younger than 50 years and the largest predicted increase in the incidence rate of colon cancer will occur in the 20- to-34-year age group. By 2020 and 2030, the incidence rate of colon cancer will increase by 37.8% and 90.0%, respectively, for patients 20 to 34 years old, while decreasing by 23.2% and 41.1%, respectively, for patients older than 50 years. For the year 2030, this represents an arithmetic difference of 131.1% in the incidence rate change of colon cancer in younger patients compared with patients older than 50 years.
The annual incidence rates of rectosigmoid and rectal cancers from 1975 to 2010 by 15-year intervals and stage of disease at diagnosis are shown in Figure 3. Since 1975, there has been a steady decline in localized, regional, and distant disease in patients aged 50 years and older. The most notable decline was observed in patients older than 75 years, for which the APCs for localized, regional, and distant disease were −1.49 (95% CI, −1.85 to −1.12; P < .001), −1.63 (95% CI, −1.97 to −1.28; P < .001), and −2.39 (95% CI, −2.63 to −2.15; P < .001), respectively. The trend observed in patients between the ages of 50 and 74 years was less dramatic, but APC is decreasing for all 3 stages of disease. The opposite trend was observed in younger patients. Quiz Ref IDThe APC is increasing for patients younger than 50 years for all 3 stages of cancer, but the increase is most significant in the 20- to 34-year age group. In the youngest age group, the APCs for localized, regional, and distant disease were 4.03 (95% CI, 3.02-5.05; P < .001), 3.05 (95% CI, 1.95-4.17; P < .001), and 2.66 (95% CI, 1.33-3.99; P < .001), respectively.
The APC-based predicted incidence rates of rectosigmoid and rectal cancers in 2020 and 2030 compared with 2010 for the 4 different age groups are shown in Figure 4. Similar to colon cancer, the incidence rates of rectosigmoid and rectal cancers will continue to decrease for patients older than 50 years, with the largest decrease occurring in patients older than 75 years. Conversely, the incidence rates of rectosigmoid and rectal cancers will continue to increase in patients younger than 50 years. As in colon cancer, the largest predicted increase in incidence rate is observed in the 20- to 34-year age group. By 2020 and 2030, the incidence rates for rectosigmoid and rectal cancers are expected to increase by 49.7% and 124.2%, respectively, for patients 20 to 34 years old, while decreasing by 23.2% and 41%, respectively, for patients older than 50 years. For the year 2030, this represents a difference of 165% in the incidence rate change of rectosigmoid and rectal cancers in younger patients compared with patients older than 50 years.
Our population-based analysis demonstrates increasing disparities in age-related incidence rates of colon and rectal cancers in the United States over the past 25 years. The incidence rates of colon and rectal cancers in young patients are increasing without corresponding increases in patients older than 50 years. Quiz Ref IDAt the present rate, the incidence rate for young patients with newly diagnosed colon or rectal cancer will nearly double by 2030, while it will similarly decline by more than one-third among patients older than the screening age of 50 years. By 2030, 10.9% of all colon and 22.9% of all rectal cancers will be diagnosed in patients younger than the screening age compared with 4.8% and 9.5%, respectively, in 2010. Stated another way, more than 1 in 10 colon cancers and nearly 1 in 4 rectal cancers will be diagnosed in people younger than the traditional screening age. These findings highlight an important growing public health problem of increasing risks for CRC among young patients.
Our finding of increasing incidence of CRC in young patients is consistent with prior studies of CRC trends using large population-based databases.2,8- 10 A 1.5%-per-year increase in the incidence of CRC in men and a 1.6%-per-year increase in women aged 20 to 49 years from 1992 to 2005 has been previously observed.8 In this study, we further provided quantitative estimates of the exponentially increasing risk for CRC among young patients compared with the declines among patients older than 50 years. Thus, our study provides further insight into the growing concern of age-related disparities in the incidence rates of CRC. Moreover, these relationships are not unique to colorectal cancer as similar findings have been observed for breast cancer, with increasing incidence among young women.11
The increasing incidence of CRC among young adults is concerning and highlights the need to investigate potential causes and external influences such as lack of screening and behavioral factors. Routine screening options for CRC include fecal occult blood test, flexible sigmoidoscopy, or colonoscopy. Adults younger than 50 years without risk factors are not recommended to undergo routine screening for CRC, whereas high-risk patients (ie, patients with a history of familial polyposis, hereditary nonpolyposis CRC, or ulcerative colitis) are recommended to begin screening prior to age 50 years.5 In the absence of routine screening, patients and health care professionals may need to have a heightened awareness regarding colorectal cancer risk and its increasing incidence to minimize delays in both diagnosis and treatment, as well as missed opportunities for treatment of precursor lesions resulting from delays in seeking medical attention. In the pre–Patient Protection and Affordable Care Act era, younger patients often lacked health insurance, resulting in further delays.
The observed increasing incidence of CRC may also be attributed to behavioral factors. Obesity, physical inactivity, and the Western diet are all risk factors associated with an increased risk for CRC. The prevalence of obesity has increased markedly in the United States and is a major risk factor for CRC.13- 15 Physical activity has been demonstrated to reduce the risk for both colon and rectal cancers.16,17 Moderate physical activity has been associated with a decreased risk for rectal cancer in both men and women, with odds ratios of 0.70 and 0.51, respectively.16 The Western diet, which consists primarily of processed meats, red meat, fast food, and low levels of vegetables and fruits, is associated with an increased risk for colon cancer.18 While the net contribution of these factors to the increasing incidence among young patients cannot be known, many of these risk factors are potentially modifiable with behavioral changes at little cost to the health care system and may impact overall health in addition to reducing the risk for CRC.
Our study demonstrates the potential impact of CRC incidence in young patients if no changes are made in public awareness and policy. Based on our data, the predicted incidence rates of colon and rectal cancers in 2030 will increase by 90% and 124%, respectively, in patients aged 20 to 34 years. Studies have demonstrated that both screening and risk factors contribute to the overall incidence of CRC. Using the MISCAN-Colon model, Edwards et al2 demonstrated that changes in risk factors and screening individually accounted for 50% of the overall decline in CRC incidence rates between 1975 and 2000. These findings demonstrated the importance of emphasizing education regarding risk factors, improving healthy lifestyles, and prevention in public policy.
The incidence rates of colon and rectal cancers are increasing in young adults and declining in adults older than 50 years. Evaluation of the predicted incidence rates reveals increases in the number of young adults who will be diagnosed as having colon or rectal cancer over the next 2 decades. Further studies are needed to determine the cause for these trends and identify potential preventive and early-detection strategies.
Corresponding Author: George J. Chang, MD, MS, Department of Surgical Oncology, University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, T Boone Pickens Academic Tower, Unit 1484, Houston, TX 77030 (firstname.lastname@example.org).
Accepted for Publication: March 6, 2014.
Published Online: November 5, 2014. doi:10.1001/jamasurg.2014.1756.
Author Contributions: Dr Chang had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: You, Skibber, Chang.
Acquisition, analysis, or interpretation of data: Bailey, Hu, Bednarski, Rodriguez-Bigas, Cantor, Chang.
Drafting of the manuscript: Bailey, You, Chang.
Critical revision of the manuscript for important intellectual content: Hu, Bednarski, Rodriguez-Bigas, Skibber, Cantor, Chang.
Statistical analysis: Bailey, Hu, Chang.
Obtained funding: Chang.
Administrative, technical, or material support: Skibber, Cantor, Chang.
Study supervision: You, Skibber, Cantor, Chang.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported in part by National Institutes of Health/National Cancer Institute grants T32CA009599 (Dr Bailey), K07-CA133187 (Dr Chang), and CA16672 (The University of Texas MD Anderson Cancer Center support grant).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Correction: This article was corrected online January 22, 2015, for an incorrect academic degree listed in the byline.